Abstract To enhance turbine efficiency, it is essential to mitigate the losses generated by irreversible phenomena in turbine flows, including boundary layers, shock waves, vortices, and trailing edge wakes. Fast and accurate detection of losses is crucial from the earliest design stages, where reduced-order models based on oversimplified correlations are employed. This requires a deep understanding of the physics behind loss-generating mechanisms, attainable by examining the 3D flow. While existing criteria can identify various phenomena, accurately quantifying vortex-generated losses remains challenging, as these losses frequently extend beyond the vortical structure. This paper provides a straightforward and effective approach to localize and assess vortex-related losses. The method is grounded in Zlatinov's decomposition of the entropy generation rate into a streamwise and secondary flow component. A criterion based on vortex kinematics evaluates the vortex's strength, enabling the determination of its spatial influence and contribution to overall losses. To validate the method, a post-processing code is developed to perform loss breakdown, using existing identification criteria and new techniques introduced within this work, particularly for wake detection. 3D RANS simulations on various configurations, from simple curved ducts to nozzle guide vanes, allow to gradually test and validate the computational tool. Results confirm that the highest entropy generation rates occur outside the vortical structure and show good ability to identify both the vortex shape and its area of influence in terms of losses. A significant improvement in predicting vortex losses is observed, especially for turbine blades with leakage vortices.
Read full abstract